William Dodds
Job Market Candidate

Stanford University
Department of Economics
579 Serra Mall
Stanford, CA 94305
510-206-1072
wdodds@stanford.edu

Curriculum Vitae

Fields:
Primary: Public Economics
Secondary: Development Economics
Expected Graduation Date:
June, 2019

Advisors:
Doug Bernheim (Co-Primary):
bernheim@stanford.edu
Caroline Hoxby (Co-Primary):
choxby@stanford.edu
Raj Chetty:
chetty@fas.harvard.edu

Job Market Paper

Sources of Income Inequality: Productivities vs. Preferences
(with Katy Bergstrom)

This paper develops a new method to understand how much of income inequality is due to differences in productivities as opposed to differences in preferences for consumption relative to leisure. In our theoretical framework, individuals have heterogeneous productivities and preferences for consumption relative to leisure. Individuals make partially observable labor supply decisions conditional on their primitives (productivities and preferences); individual optimization problems define a mapping between primitives and labor supply decisions. We show that we can invert this mapping, recovering primitives from observables. Crucially, we express this inverse mapping entirely in terms of observable reduced form elasticities with respect to the tax rate. We then implement our method empirically in the context of the U.S., showing how the relevant tax elasticities change our inferences about the sources of income inequality. Higher income effects and a larger difference between the elasticities of income and hours worked with respect to the tax rate imply that preferences play a more important role in driving income inequality. Finally, we show via simulation that once we account for both productivity and preference heterogeneity, larger income effects and larger differences between the income and hours worked elasticities imply lower optimal tax rates relative to a Mirrleesian benchmark in which all income heterogeneity is due to productivity differences.

Working Papers

The Targeting Benefit of Conditional Cash Transfers
(with Katy Bergstrom)

Conditional cash transfers (CCTs) are a popular type of social welfare program that make payments to households conditional on human capital investments in children. Compared to unconditional cash transfers (UCTs), CCTs may exclude the poorest households, as access is tied to the consumption of normal goods. However, we argue that conditionalities based on children's schooling may actually improve the targeting of transfers to low consumption households. Sending a child to school can result in a discrete loss of child income, so that schooling is negatively correlated with household consumption. Thus, schooling decisions may act as a useful "tag" for cash transfers. The size of the targeting benefit is directly related to two elasticities already popular in the literature: the income effect of a UCT and the price effect of a CCT. We estimate these elasticities for a large CCT program in rural Mexico, Progresa, using variation in transfers to younger siblings to identify income effects. We find that the targeting benefit is almost as large as the cost of excluding some low income households; this implies that if the only benefit of imposing conditions is improved targeting, 55% of the Progresa budget should go to a CCT over a UCT.

Optimal Taxation with Discontinuous Behavioral Responses

This paper explores the possibility that some individuals change their incomes by large, discrete amounts in response to small changes in the tax schedule. Variational arguments employed to solve optimal taxation problems have typically assumed away these sort of "jumping effects". While we provide an argument that shows this assumption is warranted in the standard Mirrleesian framework, we highlight two realistic extensions for which jumping effects are crucially important. First, if there is a mass of individuals who can not work, jumping effects, in the form of extensive margin responses, will occur at the bottom of the income distribution. We demonstrate through simulations that constraining the tax schedule so as to not induce jumping behavior induces welfare losses of around $3,000 per capita and leads to tax rates that are 10-40 percentage points higher than optimal. This provides a previously overlooked reason for special treatment of the extensive margin. Second, once we move to a model with multi-dimensional agent heterogeneity, we show that jumping effects occur under the optimal tax schedule in realistic cases. Accounting for jumping effects, the framework we develop allows us to extend the sufficient statistic approach to optimal taxation to cases of multi-dimensional agent heterogeneity.

Intrahousehold Investment Decisions and Child Mortality: Explaining the First Child Preference in India
(with Katy Bergstrom)

Jayachandran and Pande (2017) discovered the existence of a strong first child bias in India, suggesting that a preference for eldest sons is responsible for this birth-order-gradient. We explore whether this phenomenon could arise as a rational investment decision from the parentsí perspective. We develop a model to show that if parents are faced with significant child mortality risk and plan to rely on their children for support in old age, parents will invest more in the first child relative to their second. Our model predicts a negative relationship between investment differentials across the first- and second-born children and the infant survival rate, as well as a positive relationship between investment differentials and the early-childhood survival rate. Using data on the heights of Indian children from demographic health surveys, we find evidence that both predictions hold in the data.